Fast, Accurate and Lightweight Super-Resolution with Neural Architecture Search

TL;DR

This paper introduces an automated neural architecture search method for super-resolution that balances model simplicity and performance, outperforming many existing methods in efficiency and accuracy.

Contribution

It proposes a multi-objective neural architecture search with elastic search tactics at micro and macro levels, combining evolutionary computation and reinforcement learning.

Findings

Generated models outperform state-of-the-art methods in FLOPS efficiency

The approach achieves high super-resolution quality with lightweight models

Automated architecture search balances performance and resource constraints

Abstract

Deep convolutional neural networks demonstrate impressive results in the super-resolution domain. A series of studies concentrate on improving peak signal noise ratio (PSNR) by using much deeper layers, which are not friendly to constrained resources. Pursuing a trade-off between the restoration capacity and the simplicity of models is still non-trivial. Recent contributions are struggling to manually maximize this balance, while our work achieves the same goal automatically with neural architecture search. Specifically, we handle super-resolution with a multi-objective approach. We also propose an elastic search tactic at both micro and macro level, based on a hybrid controller that profits from evolutionary computation and reinforcement learning. Quantitative experiments help us to draw a conclusion that our generated models dominate most of the state-of-the-art methods with respect to the individual FLOPS.